/*
Christopher Blair	
Proposal Title: Do Women Make More Credible Threats? Gender Stereotypes and Crisis Bargaining

HYPOTHESES

Stated-Hyp1: "We hypothesize that female leaders in the FM and FF dyads will pay higher inconsistency costs and lower belligerence costs compared to the MM control." (p. 3)

	Test-Hyp1: Audience costs will be higher for women leaders in Female-Male dyads vs. male-male dyads, within inconsistency scenario

	Test-Hyp2: Audience costs will be higher for women leaders in Female-Male dyads vs. male-male dyads, within belligenrence scenario	
	
	Test-Hyp3: Audience costs will be higher for women leaders in female-female vs. male-male dyads, within inconsistency scenario	

	Test-Hyp4: Audience costs will be higher for women leaders in Female-Female dyads vs. male-male dyads, within belligenrence scenario	
	
Stated-Hyp2: "We also hypothesize that male leaders facing female leaders (the MF dyad) will pay
higher inconsistency costs and lower belligerence costs compared to the MM control." (p. 3)

	Test-Hyp5: Audience costs will be higher for men leaders in Male-Female dyads vs. male-male dyads, within inconsistency scenario

	Test-Hyp6: Audience costs will be higher for men leaders in male-female dyads vs. male-male dyads, within belligenrence scenario	

	
Stated-Hyp3: "we expect all of our hypothesized effects to be magnified among the respondent subgroups that are: (1) older and (2) more sexist" (p. 4)

	Test-Hyp7: Age will amplify audience costs for women leaders in Female-Male dyads vs. male-male dyads, within inconsistency scenario

	Test-Hyp8: Age will amplify audience costs for women leaders in Female-Male dyads vs. male-male dyads, within belligenrence scenario
	
	Test-Hyp9: Age will amplify audience costs for women leaders in female-female vs. male-male dyads, within inconsistency scenario

	Test-Hyp10: Age will amplify audience costs for women leaders in Female-Female dyads vs. male-male dyads, within belligenrence scenario
	
	Test-Hyp11: Age will amplify audience costs for men leaders in Male-Female dyads vs. male-male dyads, within inconsistency scenario

	Test-Hyp12: Age will amplify audience costs for men leaders in male-female dyads vs. male-male dyads, within belligenrence scenario
	
	
	Test-Hyp13: sexism will amplify audience costs for women leaders in Female-Male dyads vs. male-male dyads, within inconsistency scenario

	Test-Hyp14: sexism will amplify audience costs for women leaders in Female-Male dyads vs. male-male dyads, within belligenrence scenario
	
	Test-Hyp15: sexism will amplify audience costs for women leaders in female-female vs. male-male dyads, within inconsistency scenario

	Test-Hyp16: sexism will amplify audience costs for women leaders in Female-Female dyads vs. male-male dyads, within belligenrence scenario
	
	Test-Hyp17: sexism will amplify audience costs for men leaders in Male-Female dyads vs. male-male dyads, within inconsistency scenario

	Test-Hyp18: sexism will amplify audience costs for men leaders in male-female dyads vs. male-male dyads, within belligenrence scenario	

Stated-Hyp4: "we expect that irrespective of leader gender, Democrats will impose higher belligerence costs" (p. 4)

	Test-Hyp19: Audience costs will be higher among democrat respondents vs. republicans, within belligerence scenario
	
Stated-Hyp4: "our experiment also allows us to test whether Republicans and Democrats impose different audience costs on male and female presidents, and whether all out-partisans punish gender counterstereotypic behaviors."

	NOTE: in meeting, we decided to test the clear statement:
"our experiment also allows us to test whether Republicans and Democrats impose different audience costs on male and female presidents"

	Test-Hyp20: Republicans impose different costs on male and female presidents.	
	Test-Hyp21: Democrats impose different costs on male and female presidents.	
		
Stated-Hyp4: We plan to estimate separate models for each type of sexism to explore
differences. For instance, because benevolent sexists believe women require protection from men, these
individuals may impose higher belligerence costs on men who follow through against female initiators,
whereas hostile sexists may support belligerence in this case. (footnote 3)
	
	* Ignoring this because this was just a footnote
	
********************************************************************************
NOTES:

discussed in meeting on 3/8/2021


- We decided that the listed hypotheses are official hypotheses and the additional 
statements seem to be exploratory.

- decided to test the statement:
- "our experiment also allows us to test whether Republicans and Democrats impose different audience costs on male and female presidents"

- one moderator is sexism, measured using an 8-item scale. The main hypothesis 
is about the moderating effect of sexism, but there are additional hypotheses in 
a footnote about plans to test benevolent and hostile sexism separately. 
- We are not pursuing this because it's not a central hypothesis and doing this 
analysis adds enormously to the number of test hypotheses (i.e all 6 sexism 
hypotheses have to be re-run for benevolent and hostile sexism)

*/

clear all
use "Blair1131.dta", clear

********************************************************************************

* INDICATORS OF EXPERIMENTAL MANIPULATIONS

	tab BLAIR

	/*
	1.	Male, Republican, Eric/Steven, Stay out
	2.	Male, Republican, Eric/Steven, Not engage
	3.	Male, Republican, Eric/Steven, Engage
	4.	Male, Republican, Erica/Stephanie, Stay out
	5.	Male, Republican, Erica/Stephanie, Not engage
	6.	Male, Republican, Erica/Stephanie, Engage
	7.	Male, Democrat, Eric/Steven, Stay out
	8.	Male, Democrat, Eric/Steven, Not engage
	9.	Male, Democrat, Eric/Steven, Engage
	10.	Male, Democrat, Erica/Stephanie, Stay out
	11.	Male, Democrat, Erica/Stephanie, Not engage
	12.	Male, Democrat, Erica/Stephanie, Engage
	13.	Female, Republican, Eric/Steven, Stay out
	14.	Female, Republican, Eric/Steven, Not engage
	15.	Female, Republican, Eric/Steven, Engage
	16.	Female, Republican, Erica/Stephanie, Stay out
	17.	Female, Republican, Erica/Stephanie, Not engage
	18.	Female, Republican, Erica/Stephanie, Engage
	19.	Female, Democrat, Eric/Steven, Stay out
	20.	Female, Democrat, Eric/Steven, Not engage
	21.	Female, Democrat, Eric/Steven, Engage
	22.	Female, Democrat, Erica/Stephanie, Stay out
	23.	Female, Democrat, Erica/Stephanie, Not engage
	24.	Female, Democrat, Erica/Stephanie, Engage
	*/

	* gender of foreign leader
	recode BLAIR (1/12=0) (13/24=1), gen(foreign_woman) 
	tab foreign_woman
	
	* gender of domestic leaders
	recode BLAIR (1/3=0) (4/6=1) (7/9=0) (10/12=1) (13/15=0) (16/18=1) (19/21=0) (22/24=1), gen(domestic_woman)
	tab domestic_woman
	
	* dyad type 
	lab def dyad 0 "0 dom man/for man" 1 "1 dom man/for woman" 2 "2 dom woman/for man" 3 "3 dom woman/for woman"
	gen dyad = 0 if domestic_woman==0 & foreign_woman==0
	replace dyad = 1 if domestic_woman==0 & foreign_woman==1
	replace dyad = 2 if domestic_woman==1 & foreign_woman==0	
	replace dyad = 3 if domestic_woman==1 & foreign_woman==1
	lab val dyad dyad 
	
	tab dyad
	
	* belligerence
		gen belligerence = 1 if BLAIR==3|BLAIR==6|BLAIR==9|BLAIR==12|BLAIR==15|BLAIR==18|BLAIR==21|BLAIR==24

	* inconsistency: said will stay out but sends troops eventually
		/* NOTE: read the quex.
		There is no case where the president said will not send troops and eventually sent troops.
		There are some conditions when president said will send troops but didn't eventually send troops. We create a flag for such cases below.
		
		The following are conditions when president said will send troops:	
		BLAIR=2, 3, 5, 6, 8, 9, 11, 12, 14, 15, 17, 18, 20, 21, 23, OR 24 that if the attack continues, the United States military will push out the invaders
		
		Following are cases when president didn't eventually send troops:
		BLAIR=1, 2, 4, 5, 7, 8, 10, 11, 13, 14, 16, 17, 19, 20, 22, 23 
		*/
		gen saidyes = 1 if BLAIR==2|BLAIR==3|BLAIR==5|BLAIR==6|BLAIR==8|BLAIR==9|BLAIR==11|BLAIR==12|BLAIR==14|BLAIR==15|BLAIR==17|BLAIR==18|BLAIR==20|BLAIR==21|BLAIR==23|BLAIR==24
	
		gen didnt_send = 1 if BLAIR==1|BLAIR==2|BLAIR==4|BLAIR==5|BLAIR==7|BLAIR==8|BLAIR==10|BLAIR==11|BLAIR==13|BLAIR==14|BLAIR==16|BLAIR==17|BLAIR==19|BLAIR==20|BLAIR==22|BLAIR==22
		
		gen inconsistency = 1 if saidyes==1 & didnt_send==1
		tab inconsistency
	
* OUTCOME MEASURES

* audience costs
	//To what extent do you disapprove or approve of how the U.S. President handled the situation?
	tab Q3
 	replace Q3=. if Q3>7
	gen audiencecost=8-Q3
	tab audiencecost, mis
	
* MODERATORS

* age
	tab AGE

* sexism
/*
A.	Feminists are making reasonable demands of men. 
B.	Women must overcome more obstacles than men to be professionally successful.
C.	Women who complain about discrimination often cause more problems than they solve. // more= sexist
D.	When women demand equality these days, they are actually seeking special favors. // more= sexist
E.	Compared to men, women tend to have a superior moral sensibility. // more= sexist
F.	Many women have a quality of purity that few men possess. // more= sexist
G.	Men have no special obligation to provide financially for the women in their lives. 
H.	There is no need for men to cherish or protect women.

note that some items are reverse coded- recoding so that higher values indicate more sexism
read the citations in the proposal if there is confusion about how this scale is created
*/

* recode missing
foreach var in Q1A Q1B Q1C Q1D Q1E Q1F Q1G Q1H {
	replace `var'=. if `var'>6
	tab `var', mis
}

* recode reverse-coded iterms so that higher values indicate more sexism
foreach var in Q1A Q1B Q1G Q1H {
	gen rev`var'= 7-`var'
	tab rev`var', mis
}

	egen sexism = rowtotal(Q1A Q1B Q1C Q1D Q1E Q1F Q1G Q1H) 
	tab sexism, mis
	foreach var in Q1A Q1B Q1C Q1D Q1E Q1F Q1G Q1H {
	replace sexism=. if `var'==.
	}
	tab sexism, mis
	
	
* respondent party affiliation
	tab PARTYID7
	recode PARTYID7 (1/3=1) (5/7=0) (*=.), gen(dem_vs_rep)
	tab dem_vs_rep
	
********************************************************************************

* ANALYSIS
	

*	Test-Hyp1: Audience costs will be higher for women leaders in Female-Male dyads vs. male-male dyads, within inconsistency scenario

	reg audiencecost i.dyad if dyad==0|dyad==2 & inconsistency==1
	// do not reject. 0.000 
	tess 2.dyad +, init(Blair1131) 
	
*Test-Hyp2: Audience costs will be higher for women leaders in Female-Male dyads vs. male-male dyads, within belligenrence scenario	
	reg audiencecost i.dyad if dyad==0|dyad==2 & belligerence==1
	// reject. 0.000W
	tess 2.dyad +
	
*Test-Hyp3: Audience costs will be higher for women leaders in female-female vs. male-male dyads, within inconsistency scenario	
	reg audiencecost i.dyad if dyad==0|dyad==3 & inconsistency==1
	// do not reject. 0.000 
	tess 3.dyad +
	
*Test-Hyp4: Audience costs will be higher for women leaders in Female-Female dyads vs. male-male dyads, within belligenrence scenario	
	reg audiencecost i.dyad if dyad==0|dyad==3 & belligerence==1
	// reject. 0.000W 
	tess 3.dyad +	
	
*Test-Hyp5: Audience costs will be higher for men leaders in Male-Female dyads vs. male-male dyads, within inconsistency scenario
	reg audiencecost i.dyad if dyad==0|dyad==1 & inconsistency==1
	// do not reject. 0.000 
	tess 1.dyad +
	
*Test-Hyp6: Audience costs will be higher for men leaders in male-female dyads vs. male-male dyads, within belligenrence scenario	
	reg audiencecost i.dyad if dyad==0|dyad==1 & belligerence==1
	// reject. 0.000W 	
	tess 1.dyad +
	
*Test-Hyp7: Age will amplify audience costs for women leaders in Female-Male dyads vs. male-male dyads, within inconsistency scenario
	reg audiencecost i.dyad##c.AGE if dyad==0|dyad==2 & inconsistency==1
	// reject. 0.365 
	tess 2.dyad#AGE +
	
*Test-Hyp8: Age will amplify audience costs for women leaders in Female-Male dyads vs. male-male dyads, within belligenrence scenario	
	reg audiencecost i.dyad##c.AGE if dyad==0|dyad==2 & belligerence==1
	// reject. 0.989	
	tess 2.dyad#AGE +
	
*Test-Hyp9: Age will amplify audience costs for women leaders in female-female vs. male-male dyads, within inconsistency scenario
	reg audiencecost i.dyad##c.AGE if dyad==0|dyad==3 & inconsistency==1
	// reject. 0.896
	tess 3.dyad#AGE +

*Test-Hyp10: Age will amplify audience costs for women leaders in Female-Female dyads vs. male-male dyads, within belligenrence scenario
	reg audiencecost i.dyad##c.AGE if dyad==0|dyad==3 & belligerence==1
	// reject. 0.143 
	tess 3.dyad#AGE +	
	
*Test-Hyp11: Age will amplify audience costs for men leaders in Male-Female dyads vs. male-male dyads, within inconsistency scenario
	reg audiencecost i.dyad##c.AGE if dyad==0|dyad==1 & inconsistency==1
	// reject. 0.446
	tess 1.dyad#AGE +
	
*Test-Hyp12: Age will amplify audience costs for men leaders in male-female dyads vs. male-male dyads, within belligenrence scenario
	reg audiencecost i.dyad##c.AGE if dyad==0|dyad==1 & belligerence==1
	// reject. 0.544	
	tess 1.dyad#AGE +
	
	
*Test-Hyp13: sexism will amplify audience costs for women leaders in Female-Male dyads vs. male-male dyads, within inconsistency scenario
	reg audiencecost i.dyad##c.sexism if dyad==0|dyad==2 & inconsistency==1
	// reject. 0.065 
	tess 2.dyad#sexism +


*Test-Hyp14: sexism will amplify audience costs for women leaders in Female-Male dyads vs. male-male dyads, within belligenrence scenario
	reg audiencecost i.dyad##c.sexism if dyad==0|dyad==2 & belligerence==1
	// reject. 0.639	
	tess 2.dyad#sexism +

	
*Test-Hyp15: sexism will amplify audience costs for women leaders in female-female vs. male-male dyads, within inconsistency scenario
	reg audiencecost i.dyad##c.sexism if dyad==0|dyad==3 & inconsistency==1
	// reject. 0.455
	tess 3.dyad#sexism +
	
*Test-Hyp16: sexism will amplify audience costs for women leaders in Female-Female dyads vs. male-male dyads, within belligenrence scenario
	reg audiencecost i.dyad##c.sexism if dyad==0|dyad==3 & belligerence==1
	// reject. 0.823	
	tess 3.dyad#sexism +	
	
*Test-Hyp17: sexism will amplify audience costs for men leaders in Male-Female dyads vs. male-male dyads, within inconsistency scenario
	reg audiencecost i.dyad##c.sexism if dyad==0|dyad==1 & inconsistency==1
	// reject. 0.018W
	tess 1.dyad#sexism +
	
*Test-Hyp18: sexism will amplify audience costs for men leaders in male-female dyads vs. male-male dyads, within belligenrence scenario	
	reg audiencecost i.dyad##c.sexism if dyad==0|dyad==1 & belligerence==1
	// reject. 0.376	
	tess 1.dyad#sexism +
	
*Test-Hyp19: Audience costs will be higher among democrat respondents vs. republicans, within belligenrence scenario
	reg audiencecost i.dem_vs_rep if belligerence==1
	// do not reject. 0.000
	tess 1.dem_vs_rep +
	
*Test-Hyp20: Republicans impose different costs on male and female presidents.	
	reg audiencecost i.domestic_woman if dem_vs_rep==0
	// reject. 0.243
	tess 1.domestic_woman
	
*Test-Hyp21: Democrats impose different costs on male and female presidents.	
	reg audiencecost i.domestic_woman if dem_vs_rep==1
	// reject. 0.928
	tess 1.domestic_woman	
